The U.S. Congress has identified software reuse as an important potential money saver, to the tune of $1 billion a year, or more than 4 percent of the Defense Department's annual software development and maintenance bill.
"The '93 workshop, sponsored by the Institute of Electrical and Electronic Engineers (IEEE), revealed how far the reuse concept has come since its formalization in the early 1980s," said Jeff Poulin, WISR '93 Chairperson and Advisory Programmer of IBM Federal Systems Company, Owego, N.Y.
Current research focuses not only on "reuse-in the small," or the sharing of small utilities and functions, but on reuse in a larger sense, or the "building of programs from substantial pieces of prefabricated software," Poulin said. Several groups at the conference addressed the notion of reusing entire subsystems.
The 80 attendees covered a range of topics, including design for reuse, reuse tools and environments, language issues, reuse in education, technology transfer, certification of reusable components and management issues. Technical papers were shared and covered a wide range of topics from limits of concrete component reuse to reuse for real-time systems.
WISR '93 brought together leading reuse developers from both the public and private sectors. Participants included representatives from IBM, Hewlett-Packard, Motorola, Microsoft, Schlumberger, Boeing, Mitsubishi and Andersen Consulting. The conference also attracted researchers from the University of Maine, West Virginia University and Virginia Tech, among other educational institutions, and proponents from government organizations such as the Advanced Research Projects Agency, the Defense Information Systems Agency, the General Accounting Office and the Naval Postgraduate School.
The state of reuse education was also a concern at WISR '93, and attendees shared ideas on adding reuse courses to software engineering curriculums. Today educators emphasize new code development, typically ignoring the fact that in the business world "most programming consists of upgrading old programs or adapting someone else's work," Poulin explains.
Large-scale software reuse will require both "bottom-up" and "top-down" design strategies, both of which FSC is pursuing in different government programs. The bottom-up approach provides domain-specific toolkits, with everything a programmer needs to construct a larger system. Top-down, on the other hand, builds a high-level conceptual framework and then fills in the blanks.
FSC researchers in the Advanced Research Projects Agency's DSSA program focus on the emerging area of reuse at the software architecture level. This approach, which promises substantial reductions in software development time, is widely recognized as a key to software reuse.
According to Lloyd Mosemann, deputy assistant secretary of the Air Force, for example, standard, common software architectures "promote reusable software products that are adaptable, reliable and portable." The software industry's pursuit of this goal "reflects a maturing of the software engineering community," says Will Tracz, one of DSSA's principal investigators, FSC Owego.
The SBIS program aims at substantial reuse as well, targeting an aggressive 55 percent by the end of the first year of application development. To reach this goal, Owego will use both bottom-up and top-down approaches. FSC will build a standard collection of data types, standard interfaces to all commercial products bid on the program, standard application programming interfaces and a domain architecture that will help applications developers to fit the pieces together to construct larger systems.
IBM FSC in Gaithersburg is a prime contractor in the ARPA STARS (Software Technology for Adaptable Reliable Systems) program which over the past five years has pioneered many of the modern reuse technologies. One of IBM's responsibilities in STARS is the operation of ASSET (Asset Source for Software Engineering Technology), an operational software reuse library intended to help foster a reuse industry.